Google ACE Practice Test 32 — 15 Questions

Question 1

After deploying a Kubernetes Deployment named 'web-app', a developer wants to expose it externally on a static IP address. Which kubectl command should they use?

Accepted Answer

kubectl expose deployment web-app --type=LoadBalancer --port=80 --target-port=8080. kubectl expose creates a service. To get an external load balancer with a static IP, the service type must be LoadBalancer. The --type flag specifies the service type.

Answer

kubectl create service clusterip web-app --tcp=80:8080

Answer

gcloud compute forwarding-rules create web-app --port=80

Answer

kubectl expose deployment web-app --type=NodePort --port=80

Question 2

A team is deploying a microservice to Cloud Run that needs to process messages from Pub/Sub. The service should only be invocable by Pub/Sub push deliveries, not by unauthenticated HTTP requests. What should the team do?

Accepted Answer

Deploy with --no-allow-unauthenticated and configure the Pub/Sub subscription to use a service account that has the roles/run.invoker role on the Cloud Run service. To restrict invocation to only Pub/Sub, the Cloud Run service must require authentication and the Pub/Sub subscription must be configured to use a service account to push. The --no-allow-unauthenticated flag ensures only authenticated requests are accepted, and the Pub/Sub subscription's push endpoint must be set with the service's URL and use a service account with the run.invoker role.

Answer

Deploy with --allow-unauthenticated and set up a Pub/Sub subscription with OIDC token audience

Answer

Deploy with --no-allow-unauthenticated and create a VPC connector to allow Pub/Sub internal traffic

Answer

Use Cloud Functions instead, which is more secure for Pub/Sub triggers

Question 3

A team is using Terraform to manage Google Cloud infrastructure. They want to store Terraform state files in a remote backend that supports locking to prevent concurrent modifications. Which backend should they use?

Accepted Answer

Cloud Storage. Google Cloud Storage (GCS) is the recommended backend for Terraform on GCP. It supports state locking via object versioning and a lock mechanism (using the state lock table in a GCS bucket). Cloud Shell is not a backend. Cloud Source Repositories is for source code. Bigtable is for NoSQL workloads.

Answer

Cloud Source Repositories

Answer

Cloud Shell

Answer

Cloud Bigtable

Question 4

You need to update the container image of a deployment named 'my-app' in GKE to a new version. Which command should you use?

Accepted Answer

kubectl set image deployment/my-app my-app-container=my-image:v2. kubectl set image updates the image of a deployment.

Answer

kubectl apply -f updated-deployment.yaml

Answer

kubectl update deployment my-app --image=my-image:v2

Answer

kubectl edit deployment my-app --image=my-image:v2

Question 5

You are deploying a new revision of a Cloud Run service. You want to gradually shift traffic from the old revision to the new one, starting with 10% traffic to the new revision. Which command should you use?

Accepted Answer

gcloud run services update-traffic my-service --to-revision new-revision --percent 10. Cloud Run allows traffic splitting between revisions. The gcloud run services update-traffic command can specify the percentage of traffic for each revision.

Answer

gcloud run revisions traffic my-service --to-revision new-revision --percent 10

Answer

gcloud run services update my-service --image gcr.io/my-project/my-image:new

Answer

gcloud run deploy my-service --image gcr.io/my-project/my-image:new --traffic new-revision=10

Question 6

You notice that a deployment in your GKE cluster is running an outdated image. You need to update the deployment to use the new image 'gcr.io/my-project/my-app:v2'. Which kubectl command should you use?

Accepted Answer

kubectl set image deployment/my-deployment my-app=gcr.io/my-project/my-app:v2. To update the image of a deployment, use 'kubectl set image' specifying the deployment name and the container name:tag.

Answer

kubectl rollout restart deployment my-deployment --image gcr.io/my-project/my-app:v2

Answer

kubectl update deployment my-deployment --image gcr.io/my-project/my-app:v2

Answer

kubectl replace deployment my-deployment --image gcr.io/my-project/my-app:v2

Question 7

A company wants to run a stateful application on Compute Engine with persistent storage that can be attached to another instance in case of failure. Which storage option should they use?

Accepted Answer

Persistent Disk. Persistent Disk (PD) is a network-attached block storage that can be detached and reattached to another VM in the same zone. Option C is correct. Option A (Filestore) is a file storage service, not block storage; Option B (Cloud Storage bucket) is object storage; Option D (Local SSD) is ephemeral and tied to the instance.

Answer

Filestore

Answer

Cloud Storage bucket

Answer

Local SSD

Question 8

A site reliability engineer needs to be notified immediately when the error rate of a production microservice exceeds 5% over a 5-minute window. Which type of alerting policy should be used?

Accepted Answer

Metric threshold alert. A metric threshold alert triggers when a metric crosses a threshold. This scenario requires tracking the error rate metric and alerting when it exceeds 5%.

Answer

Uptime check alert

Answer

Pub/Sub notification hook

Answer

Log-based alert (log metric trigger)

Question 9

You have a Compute Engine VM that is running a critical application. You need to change its machine type from n1-standard-4 to n2-standard-8. What is the correct procedure?

Accepted Answer

Stop the instance, then use gcloud compute instances set-machine-type, then start the instance. To change the machine type of a VM, you must stop the instance first, then use the gcloud compute instances set-machine-type command, and finally start the instance again.

Answer

Use gcloud compute instances update --machine-type n2-standard-8 while the instance is running

Answer

Delete the instance and create a new one with the desired machine type

Answer

Use gcloud compute instances resize --machine-type n2-standard-8 without stopping

Question 10

You want to monitor the uptime of an external HTTP endpoint every minute and receive an email notification if the endpoint is unavailable for more than two consecutive checks. What should you do?

Accepted Answer

Create an uptime check in Cloud Monitoring, then create an alerting policy with condition 'metric threshold' for 'check_failed' and set notification channel to email. Uptime checks in Cloud Monitoring can be configured to check HTTP endpoints. You can set alerting conditions based on the duration of the outage and choose email as a notification channel.

Answer

Create a log-based alert in Cloud Logging that triggers on network errors

Answer

Use Cloud Functions to periodically call the endpoint and send an email on failure

Answer

Configure a TCP health check on the load balancer

Question 11

You need to drain a GKE node for maintenance. The node is running a DaemonSet and some pods with emptyDir volumes. Which kubectl command should you use to safely drain the node without causing errors?

Accepted Answer

kubectl drain node-name --ignore-daemonsets --delete-emptydir-data. The kubectl drain command with --ignore-daemonsets and --delete-emptydir-data flags safely evicts pods while ignoring DaemonSets (which are managed by the node) and allowing deletion of pods with emptyDir volumes.

Answer

kubectl drain node-name --force

Answer

kubectl drain node-name --ignore-daemonsets

Answer

kubectl cordon node-name && kubectl delete pods --all --grace-period=0

Question 12

You need to be notified when the CPU utilization of any Compute Engine instance in your project exceeds 80% for 5 minutes. Which Cloud Monitoring feature should you use?

Accepted Answer

Metric threshold alerting policy. Metric threshold alerting policies allow you to set conditions based on metric values. When the condition (CPU > 80% for 5 minutes) is met, the alert fires and sends notifications via configured channels.

Answer

Uptime check

Answer

Log-based alert

Answer

Dashboard

Question 13

Your BigQuery query is taking longer than expected. You want to estimate the query cost before running it and get a preview of how many bytes will be processed. Which bq command should you use?

Accepted Answer

bq query --use_legacy_sql=false --dry_run 'SELECT ...'. The bq query command with the --dry_run flag (or --dry-run) will process the query and return the amount of data that would be scanned, without executing the query. This helps estimate cost and performance.

Answer

bq show --format=prettyjson mydataset.mytable

Answer

bq ls --format=prettyjson mydataset

Answer

bq query --use_legacy_sql=false --batch 'SELECT ...'

Question 14

Your team uses Cloud Logging to store application logs. You want to create a metric that counts the number of ERROR log entries per service. Which type of log-based metric should you create?

Accepted Answer

Counter metric. Log-based metrics can be counter metrics (count of log entries matching a filter) or distribution metrics. For counting occurrences, a counter metric is appropriate.

Answer

Distribution metric

Answer

Boolean metric

Answer

Gauge metric

Question 15

You are deploying a GKE cluster with node autoscaling enabled. The cluster runs batch jobs that are sensitive to startup latency. You notice that during scale-up, new nodes take several minutes to become ready. Which action can reduce the time it takes for new nodes to join the cluster?

Accepted Answer

Use a custom image with pre-installed dependencies. Using a custom image with pre-installed dependencies reduces the time needed for node initialization because the image already contains the required software, avoiding downloads during startup. This is especially beneficial for batch jobs.

Answer

Increase the initial node pool size

Answer

Set the --max-nodes-per-pool flag to a higher value

Answer

Enable cluster autoscaler with --enable-autorepair